1)Define a perceptron. Discuss the perceptron algorithm to update the weights.

The perceptron is a mathematical model of a biological neuron.It's a model of a single neuron

that can be used for binary classification problems. This means it can decide if an input belongs
to one class or another.

Algorithm-class notes

2)What do you mean by Hebbian Learning? Explain the Hebb’s Learning algorithm

Hebbian Learning Rule, also known as Hebb Learning Rule, was proposed by Donald O Hebb. It
is one of the first and also easiest learning rules in the neural network. It is used for pattern
classification. It is a single layer neural network, i.e. it has one input layer and one output layer.

Hebbian Learning Rule Algorithm :

Set all weights to zero, wi = 0 for i=1 to n, and bias to zero.

For each input vector, S(input vector) : t(target output pair), repeat steps 3-5.

Set activations for input units with the input vector Xi = Si for i = 1 to n.

Set the corresponding output value to the output neuron, i.e. y = t.

Update weight and bias by applying Hebb rule for all i = 1 to n:

3)With neat diagrams, explain the concept McCulloch Pitts Neuron Model. Illustrate how it
is helpful in handling the Neural Network problems

Mcculloch pitts neuron model:

This is simplified model of real neurons, known as Threshold Logic Unit.

A set of synapsesc (i.e connections) brings the activations from the other neurons.
A processing unit sums the inputs, the applies the non-linear activation funcation

Components of the McCulloch-Pitts Neuron Model:

Inputs (xi): The neuron receives binary inputs (usually 0 or 1) from other neurons.These inputs
represent the activation signals.

Weights (wi): Each input has an associated weight (wi) that determines the strength of its
influence on the neuron.

Weighted Sum: The weighted sum (S) of inputs is calculated as the sum of the products of
inputs and their corresponding weights.

Threshold (θ): The neuron has a threshold value (θ). If the weighted sum exceeds this threshold,
the neuron fires and produces an output of 1; otherwise, it remains inactive with an output of 0.
Usefulness in Handling Neural Network Problems:

Basic Building Block: The M-P neuron model serves as a foundational building block for
understanding more complex neural network architectures.

Learning Mechanism: it inspired the development of learning algorithms for adjusting weights
based on input patterns and desired outputs.

Representation of Logic Gates: This highlights its utility in solving problems related to logic
and computation.

Conceptual Framework: The M-P neuron model provides a conceptual framework for
understanding how information processing can occur in a distributed and interconnected manner.

4)List out various applications of Artificial Neural Networks with illustrations

➔ Image Recognition:

ANNs can be trained to recognize objects in images. For example, in facial recognition, ANNs
can identify and distinguish faces in a photograph.

➔ Speech Recognition:

ANNs are used in speech recognition systems, enabling devices like smartphones to understand
and interpret spoken language.

➔ Autonomous Vehicles:

ANNs play a crucial role in enabling self-driving cars to recognize and respond to various
objects, pedestrians, and road conditions.

➔ Medical Diagnosis:

ANNs can analyze medical data such as X-rays, MRIs, and patient records to assist in diagnosing
diseases or conditions, like identifying cancerous tumors.

➔ Financial Forecasting:

ANNs can analyze historical financial data to predict stock prices, currency exchange rates, and
other market trends.

➔ Fraud Detection:

ANNs are employed in detecting fraudulent activities in banking and financial transactions by
learning patterns indicative of fraudulent behavior.
 ➔ Recommendation Systems:

ANNs power recommendation algorithms on platforms like Netflix and Amazon, suggesting
movies, products, or content based on user preferences and behavior.

➔ Cybersecurity:

ANNs can identify patterns indicative of cyber attacks, helping in the early detection and
prevention of security breaches.

5)What do you mean by Feed forward Networks? Differentiate them from Feedback
Networks

Feed-forward neural networks transmit data in one direction—from input to output—without

feedback loops, making them suitable for tasks like pattern recognition and classification.

Feedback Control System Feed Forward Control System

In feedback system output depends on the In feed forward system the signal is passed to
generated feedback signal. some external load.
Measure of disturbances in the system is not Measure of disturbances in the system is
needed by feedback system. needed by feed forward system.
All the disturbances are detected in feedback All the disturbances are not detected in feed
system. forward system.
The loop in a feedback control system is a The loop in a feed forward control system is an
closed loop. open loop.
It focuses on the output of the system. It focuses on the input of the system.
The variables are adjusted on the basis of The variables are adjusted on the basis of
errors. knowledge

6) Write Short Notes on the following

• Weights
• Activation Function
• Learning Rate
• Bias
• Threshold
OR
Explain the Concept Artificial Neural Network? With neat diagrams explain the structure,
various types of Learnings and Activation functions in ANN

An Artificial Neural Network (ANN) is a machine learning model that mimics the way nerve
cells work in the human brain. ANNs are made up of interconnected nodes (neurons) that are
organized into layers. Data flows through these nodes, adjusting the weights of connections to
learn patterns and make predictions.
• Weights determine how much influence each input has on the output of the neuron
• The threshold value is applied to the output of the neuron before the activation function is
applied. If the output of the neuron is greater than or equal to the threshold value, then the
neuron is activated.
• Bias in a neural network is a constant value that is added to the weighted sum of the
inputs to a neuron.
• An activation function is a mathematical function that is applied to the weighted sum of
the inputs to a neuron in a neural network. It determines whether the neuron should be
activated or not.

Learning Rate:
Denoted as “a”(alpha).
Used to control the weight adjustment at each step of training.
Ranges from 0-1, and determines at each time

The learning rate is a hyperparameter that controls how quickly the weights of a neural
network are updated during training. A higher learning rate will result in faster updates to
the weights. A lower learning rate will result in slower updates to the weights